=Paper=
{{Paper
|id=None
|storemode=property
|title=GeoTriples: a Tool for Publishing Geospatial Data as RDF Graphs Using R2RML Mappings
|pdfUrl=https://ceur-ws.org/Vol-1272/paper_117.pdf
|volume=Vol-1272
|dblpUrl=https://dblp.org/rec/conf/semweb/KyzirakosVSMK14
}}
==GeoTriples: a Tool for Publishing Geospatial Data as RDF Graphs Using R2RML Mappings==
https://ceur-ws.org/Vol-1272/paper_117.pdf
GeoTriples: a Tool for Publishing Geospatial
Data as RDF Graphs Using R2RML Mappings
Kostis Kyzirakos1 , Ioannis Vlachopoulos2 , Dimitrianos Savva2 ,
Stefan Manegold1 , and Manolis Koubarakis2
1
Centrum Wiskunde & Informatica, Amsterdam, The Netherlands
{firstname.lastname}@cwi.nl
2
National and Kapodistrian University of Athens, Greece
{johnvl,dimis,koubarak}@di.uoa.gr
Abstract. In this paper we present the tool GeoTriples that allows the
transformation of Earth Observation data and geospatial data into RDF
graphs, by using and extending the R2RML mapping language to be able
to deal with the specificities of geospatial data. GeoTriples is a semi-
automated tool that transforms geospatial information into RDF follow-
ing the state of the art vocabularies like GeoSPARQL and stSPARQL,
but at the same time it is not tightly coupled to a specific vocabulary.
Keywords: Linked Geospatial Data, data publishing, GeoSPARQL, stSPARQL
1 Introduction
In the last few years there has been significant effort on publishing EO and
geospatial data sources as linked open data. However, the problem of publishing
geospatial data sources into RDF graphs using a generic and extensible frame-
work has received little attention as it has only recently emerged. Instead, script-
ing methods, that were adapted to the subject, were employed mostly for this
task, such as custom python scripts developed in project TELEIOS3 . However,
some work towards developing automated methods for translating geospatial
data into RDF has been presented in the latest LGD Workshop4 . In this paper
we present the tool GeoTriples that allows the transformation of geospatial data
stored in spatially-enabled relational databases and raw files. It is implemented
as an extension to the D2RQ platform5 [1] and goes beyond the state of the
art by extending the R2RML mapping language6 to deal with the specifities of
geospatial data. GeoTriples uses GeoSPARQL7 as the target vocabulary but the
user is free to use any vocabulary she finds appropriate.
3
http://www.earthobservatory.eu
4
http://www.w3.org/2014/03/lgd
5
http://d2rq.org
6
http://www.w3.org/TR/r2rml/
7
http://www.opengeospatial.org/standards/geosparql/
osm_w:1 rdf:type geo:Feature ;
NAME TYPE WIDTH osm_ont:hasName "Mangfall"^^xsd:string ;
Zeitlbach stream 1 geo:hasGeometry osm_g:1 .
Mangfall river 25 osm_g:1 rdf:type geo:Geometry ;
Triftbach canal 10 geo:dimension "2"^^xsd:integer .
(a) Example data from an ESRI shapefile (b) Expected RDF triples about Mangfall
_:osm
rr:logicalTable [ rr:tableName "‘osm‘"; ];
rr:subjectMap [
rr:class geo:Feature;
_:osmGeometry
rr:template "http://data.example.com/osm-waterways/
rr:logicalTable [ rr:tableName "‘osm‘"; ];
Feature/id/{‘gid‘}"; ];
rr:subjectMap [
rr:predicateObjectMap [
rr:class geo:Geometry;
rr:predicate osm:hasName; rr:template "http://data.example.com/osm-waterways/
rr:objectMap [ rr:datatype xsd:string; Geometry/id/{‘gid‘}"; ];
rr:column "‘NAME‘"; ]; ]; rr:predicateObjectMap [
rr:predicateObjectMap [ rr:predicate geo:dimension;
rr:predicate geo:hasGeometry ; rr:objectMap [
rr:objectMap [ rrx:transformation [
rr:parentTriplesMap _:osmGeometry; rrx:function geof:dimension;
rr:joinCondition [ rrx:argumentMap (
rr:child "gid"; [rr:column "‘geom‘"]
rr:parent "gid"; ]; ]; ]. ); ] ]; ].
(c) Mapping of thematic information (d) Mapping of geometric information
Fig. 1: Examples of extended R2RML mappings for OSM
2 The Tool GeoTriples
GeoTriples8 is an open source tool, that takes as input geospatial data that are
stored in a spatially enabled database, data that reside in raw files (e.g. ESRI
shapefiles) or the results that derive from processing of the aforementioned data
(e.g. a SciQL query over raster or array data). At a lower level, GeoTriples uses a
connector for each type of input data that transparently accesses and processes
the input data. It also consists of two main components: the mapping generator
and the R2RML processor. The mapping generator creates automatically an
R2RML mapping document from the input data source. The mapping is also
enriched with subject and predicate object maps so that the RDF graph that
will be produced follows the GeoSPARQL vocabulary. Geospatial information is
modeled using a variety of data models (e.g., relational, hierarchical) and is made
available using a variety of formats (e.g., ESRI shapefiles, KML documents). In
order to deal with these specificities of geospatial information, we extended the
R2RML language to allow the representation of a transformation function over
the input data via an object map. In [2] we provide more information about our
approach. Figure 1 presents an example of such a transformation. The R2RML
processor is responsible for producing the desired RDF output by taking into
account the mapping document generated, which is also optionally edited by the
user. When the R2RML processor of GeoTriples detects an object map with a
transformation function, it applies on the fly this function on the serialization
of the geometry described in the subject map. However, if the input data source
is a spatially enabled database, it generates the appropriate SQL queries that
push these transformations to the underlying DBMS.
8
https://sourceforge.net/projects/geotriples/
Fig. 2: The graphical user interface of GeoTriples
3 Using GeoTriples in a real-world scenario
In this section we present how we will demonstrate the tool GeoTriples in the
context of a precision farming application that is developed by the FP7 EU
project LEO9 . The application combines traditional geospatial data with linked
geospatial data for enhancing the quality of precision farming activities. Precision
farming aims to solve numerous problems for farmers such as the minimization
of the environmental pollution by fertilizers. For dealing with this issue, the
farmers have to comply with many legal and technical guidelines that require
the combination of information that resides in diverse information sources. In
this section we present how linked geospatial data can form the knowledge base
for providing solutions for this problem. We will publish the following datasets
as RDF graphs using GeoTriples in order to use them in the precision farming
application.
OpenStreetMap (OSM) is a collaborative project for publishing free maps
of the world. OSM maintains a community-driven global editable map that gath-
ers map data in a crowdsourcing fashion.
Talking Fields aims to increase the efficiency of agricultural production
via precision farming by means of geo-information services integrating space
and ground-based assets. It produces products for improved soil probing using
satellite-based zone maps, and provide services for monitoring crop development
through provision of biomass maps and yield estimates.
Natura 2000 is a European ecological network where national authorities
submit a standard data form that describes each site and its ecology in order to
be characterized as a Natura site.
Corine Land Cover (CLC) is an activity of the European Environment
Agency that collects data regarding the land cover of European countries.
In this demo we will use GeoTriples in order to produce the R2RML map-
pings that dictate the process of generating the desired RDF output from the
above data. Then, using the R2RML processor of GeoTriples, we translate the
input data into RDF graphs and store the latter into the geospatial RDF store
Strabon10 [3].
9
http://www.linkedeodata.eu
10
http://www.strabon.di.uoa.gr/
SELECT DISTINCT ?field_name ?river_name
WHERE {?river rdf:type osmo:River;
osmo:hasName ?river_name;
geo:hasGeometry ?river_geo .
?river_geo geo:asWKT ?river_geowkt .
?field rdf:type tf:Field;
tfo:hasFieldName ?field_name;
tf:hasRasterCell ?cell .
?cell geo:hasGeometry ?cell_geo ;
?cell_geo geo:asWKT ?field_geowkt .
FILTER(geof:distance(?river_geowkt,
?field_geowkt, uom:meter)<100)}
(a) GeoSPARQL Query (b) Query results
Fig. 3: Discover the parts of the agricultural fields that are close to rivers.
The user can use the graphical interface of GeoTriples that is displayed in
Figure 2 for publishing these datasets as RDF graphs. At first the user defines
the necessary credentials of the DBMS that stores the above datasets. Then, she
selects the tables and the columns that contain the information that she want to
publish as RDF graphs. Optionally, an existing ontology may be loaded, in order
to map the columns of the selected table to properties from the loaded ontology
and map the instances generated from the rows to a specific class. Afterwards,
GeoTriples generates automatically the R2RML mappings and presents them
to the user. Finally, the user may either customize the generated mappings or
proceed with the generation of the RDF graph.
A series of GeoSPARQL queries will be issued afterwards in Strabon for
providing the precision farming application with information like the location of
agricultural fields that are close to a river. This information allows the precision
farming application to take into account legal restrictions regarding distance
requirements when preparing the prescription maps that the farmers will utilize
afterwards. In Figure 3a we present a GeoSPARQL query that discovers this
information, and in Figure 3b we depict the query results.
4 Conclusions
In this paper we presented the tool GeoTriples that uses an extended form
of R2RML mapping language to transform geospatial data into RDF graphs,
and the GeoSPARQL ontology to properly express it. We demonstrate how
GeoTriples is being used for publishing geospatial information that resides in
different data sources for the realization of a precision farming application.
References
1. C. Bizer and A. Seaborne. D2RQ-treating non-RDF databases as virtual RDF
graphs. In Proceedings of the 3rd International Semantic Web Conference, 2004.
2. K. Kostis, V. Ioannis, S. Dimitrianos, M. Stefan, and K. Manolis. Data models and
languages for mapping EO data to RDF. Del. 2.1, FP7 project LEO, 2014.
3. K. Kyzirakos, M. Karpathiotakis, and M. Koubarakis. Strabon: A Semantic Geospa-
tial DBMS. In International Semantic Web Conference, 2012.